The major progesterone-modulated proteins secreted into the sheep uterus are members of the serpin superfamily of serine protease inhibitors

The uterine milk proteins (UTMP) are a pair of structurally related basic glycoproteins that when newly synthesized carry phosphorylated mannosyl residues on their carbohydrate chains. They are the major proteins secreted by ovine endometrium under the influence of progesterone. RNA from a late pregnant ewe endometrium was isolated for use in in vitro translation assays and for constructing cDNA libraries. Translation experiments, initially with total cellular RNA and subsequently with RNA selected by hybridization with a specific cDNA, demonstrated the production of two polypeptides (Mr = 47,000 and 55,000) that were precipitated with antiserum to the UTMP. With microsomal membranes in the translation assay, there was increased production of an Mr = 57,000 form that was protected from protease digestion. Antibody screening of a cDNA library in lambda gt11 identified a short clone representing the 3' terminus of the mRNA that was shown by epitope selection experiments to be UTMP specific. This clone was then used to screen a lambda gt10 library. A longer clone (1.3 kilobases) was isolated and sequenced but lacked the 5' terminus to the mRNA. The latter sequence was obtained directly from the mRNA. Interesting features of the UTMP mRNA sequence, which was 1,352 bases long and contained a 1,287-base open reading frame, were two strong start codons, two potential sites for N-glycosylation and a repeat of 21 bases, six bases apart, that resulted in a repeat of seven amino acids. The inferred amino acid sequence agreed closely with the NH2-terminal amino acid sequence obtained directly from the UTMP and clearly placed the UTMP in the serpin superfamily of protease inhibitors. However, we have been unable to demonstrate inhibitory activity toward any serine protease so far tested.     

Intron structure of the human antithrombin III gene differs from that of other members of the serine protease inhibitor superfamily

Antithrombin III (ATIII) plays an integral role in the coagulation system by inhibiting thrombin and several other activated clotting factors. Inherited deficiency of ATIII is quite common and can result in life-threatening thrombotic complications. In order to understand the basis of ATIII deficiency, we have isolated and characterized the normal human ATIII gene from a recombinant Charon 4A bacteriophage genomic library. The ATIII gene contains six exons and five introns distributed over approximately 19 kilobases of DNA. The positions of introns in the ATIII gene were compared with other members of the serine protease inhibitor family which share 17-31% amino acid homology. When aligned to achieve maximal protein homology, only one of the ATIII introns corresponded to the four introns of rat angiotensinogen or human alpha 1-antitrypsin. Similarly, only one ATIII intron was homologous to the seven introns of chicken ovalbumin. We present two testable models to explain the discrepancy in intron positions among members of the serine protease inhibitor superfamily of genes.     

Characterization of a cDNA for human protein C inhibitor. A new member of the plasma serine protease inhibitor superfamily

A cDNA library in lambda-phage lambda gt11 containing DNA inserts prepared from human liver mRNA was screened with monoclonal antibodies to human protein C inhibitor. Six positive clones were isolated from 6 X 10(6) phages and plaque purified. The cDNA in the phage containing the largest insert, which hybridized to a DNA probe prepared on the basis of the amino-terminal amino acid sequence of the mature inhibitor, was sequenced. This cDNA insert contained 2106 base pairs coding for a 5'-noncoding region, a 19-amino acid signal peptide, a 387-amino acid mature protein, a stop codon, and a long 3'-noncoding region of 839 base pairs. Based on the amino acid sequence of the carboxyl-terminal peptide released by cleavage of protein C inhibitor by activated protein C as well as by thrombin, the reactive site peptide bond of protein C inhibitor is Arg354-Ser355. Five potential carbohydrate-binding sites were found in the mature protein. The high homology of the amino acid sequence of protein C inhibitor to the other known inhibitors clearly demonstrates that protein C inhibitor is a member of the superfamily of serine protease inhibitors including alpha 1-antichymotrypsin, alpha 1-antitrypsin, antithrombin III, ovalbumin, and angiotensinogen. Based on the difference matrices for these proteins, we present possible phylogenetic trees for these proteins.     

Evidence that peroxiredoxins are novel members of the thioredoxin fold superfamily.

Peroxiredoxins catalyze reduction of hydrogen peroxide or alkyl peroxide, to water or the corresponding alcohol. Detailed analysis of their sequences indicates that these enzymes possess a thioredoxin (Trx)-like fold and consequently are homologues of both thioredoxin and glutathione peroxidase (GPx). Sequence- and structure-based multiple sequence alignments indicate that the peroxiredoxin active site cysteine and GPx active site selenocysteine are structurally equivalent. Homologous peroxiredoxin and GPx enzymes are predicted to catalyze equivalent reactions via similar reaction intermediates.     

Identification and preliminary characterization of two human digitalis-like substances that are structurally related to digoxin and ouabain.

In order to characterize the structure of endogenous digitalis-like immunoreactive factor (DLIF), we utilized peritoneal dialysis fluid from patients with chronic renal failure as a source of endogenous digitalis-like immunoreactive factor (DLIF), and subjected it to one-step ion exchange chromatography, followed by one step reverse HPLC. Crude dialysis fluid contained 0.09 ng/ml of DLIF, and using Amberlite XAD-2 chromatography we extracted 110 ng of DLIF from 800 ml of dialysis fluid. By applying this partially purified DLIF to our HPLC system, we discerned three peaks of DLIF activity, with retention times of 34, 58 and 63 minutes. The first peak overlapped the elution profile of ouabain, and the third peak co-eluted precisely with digoxin. The second DLIF peak was not in proximity to any of the digitalis-like markers employed. Thus, our results indicate that DLIF isolated from peritoneal dialysis fluid exists in three distinct forms, one of which resembles ouabain, and one which is identical to digoxin.     

Identification of two novel members of erbA superfamily by molecular cloning: the gene products of the two are highly related to each other.

Two v-erbA-related genes, named ear-2 and ear-3, have been identified in the human genome and characterized by cDNA cloning. These genes are predicted to encode proteins that are very similar in primary structure to receptors for steroid hormones or thyroid hormone (T3). In addition, amino acid sequences of the ear-2 and ear-3 gene products are very similar each other especially at the DNA binding domain (86% homology) and at the putative ligand binding domain (76% homology). Northern hybridization with ear DNA probes of RNAs from various tissues of a human fetus reveals that the expression of ear-2 is high in the liver whereas the expression of ear-3 is relatively ubiquitous. Hybridization analysis of DNAs from sorted chromosomes shows that the ear-2 gene is located on chromosome 19 and ear-3 on chromosome 5, indicating that the two genes are clearly different from each other.     

The two proteins of the erythropoietin receptor are structurally similar

The structure of the erythropoietin receptor has been identified in this laboratory as two proteins of 100 and 85 kDa by cross-linking 125I-erythropoietin (125I-EP) to the surface of erythroid cells purified from the spleens of mice infected with the anemia strain of Friend virus. This study investigates the relatedness of these two proteins and the possibility that these proteins are subunits of the functional receptor for EP. Other workers have claimed that the 100- and 85-kDa proteins are bridged by disulfide bonds. This most likely is an artifact due to the insolubility of the cross-linked membrane. Proteolytic digestion by the method of Cleveland (Cleveland, D. W., Fischer, S. G., Kirschner, M. W., and Laemmli, U. K. (1977) J. Biol. Chem. 252, 1102-1106) resulted in identical fragments from the 100- and 85-kDa proteins, which strongly suggests that the primary amino acid sequence of these two proteins is similar if not identical. Increasing the number of protease inhibitors during the preparation of membranes and the binding and cross-linking steps increased the ratio of 100-kDa protein labeled compared to the 85-kDa protein. Together these results suggest that the 85-kDa protein is derived by proteolytic cleavage of the 100-kDa receptor for EP. It is not clear whether the 100-kDa protein can bind EP in the absence of the 85-kDa protein.     

Characterization of proteins structurally related to human N-acetyl-beta-D-glucosaminidase.

Those proteins of human liver that cross-reacted with antibodies raised to apparently homogenous hexosamindases A and B were detected by immunodiffusion. Cross-reacting proteins with high molecular weights (greater than 2000000) and intermediate molecular weights (70000--200000) were present both in the unadsorbed fraction and in the 0.05--0.2M-NaCl eluate obtained by DEAE-cellulose chromatography at pH7.0. The unadsorbed fraction also contained a cross-reacting protein of low molecular weight (10000--70000). The possible structural and functional relationships between hexosaminidase and the cross-reacting proteins are discussed. An apparently cross-reacting protein present in the 0.05M-NaCl eluate from the DEAE-cellulose column was serologically unrelated to hexosaminidase, but it gave a reaction of immunological identify with one of the apparently cross-reacting proteins having the charge and size characteristics of hexosaminidase A. It is suggested that immunochemical methods may provide criteria for the homogeneity of enzyme preparations superior to those of conventional methods.     

Higher eucaryotic cdc25 proteins are structurally related to phosphoseryl/threonyl protein phosphatases.

cdc25 proteins are universally involved in the control of cell division. Using an original method of sequence analysis, cdc25 proteins from different sources were compared to protein phosphatases. Protein phosphatases could clearly be characterized as two distinct protein families, the phospho-seryl/threonyl phosphatases, and the phospho-tyrosyl phosphatases. None of the cdc25 proteins analyzed fitted with the phospho-tyrosyl phosphatases, indicating that if they indeed possess this biochemical activity, they form a distinct phsophatase protein group. Unexpectedly, higher eucaryotic cdc25 proteins (from human and fly) were found to be structurally related to phospho-seryl/threonyl phosphatases. These results fit well with expected function of the proteins, associated solely in higher eucaryotes, to dephosphorylation of threonine in the cell cycle control protein cdc2.     

Vaccinia virus encodes an active thymidylate kinase that complements a cdc8 mutant of Saccharomyces cerevisiae.

A vaccinia virus open reading frame (ORF) previously predicted to encode thymidylate kinase (TmpK) is shown to encode an active enzyme. A copy of the ORF, generated by polymerase chain reaction, was cloned into an Escherichia coli inducible expression vector. Cell extracts of E. coli expressing the vaccinia gene contained high levels of TmpK activity, whereas extracts of cells without the TmpK gene did not. The vaccinia ORF expressed from a yeast vector complemented a Saccharomyces cerevisiae cdc8 mutant, demonstrating functional compatibility of the vaccinia virus and yeast TmpK enzymes. The gene is shown to be nonessential for the replication of vaccinia virus in cultured cells by the construction of a viable virus mutant that has the coding region of the TmpK gene interrupted by the Ecogpt gene. Synthesis of the vaccinia TmpK protein in infected cells was demonstrated by the use of a polyvalent rabbit antiserum raised against the purified TmpK enzyme expressed in E. coli to immunoprecipitate a 23-kDa early polypeptide from cells infected with wild type vaccinia but not from cells infected with the TmpK mutant. Plasmid vectors that allow the construction of recombinant viruses expressing foreign gene(s) from the nonessential TmpK locus are described.     

Inter-alpha-trypsin inhibitor: emergence of a family within the Kunitz-type protease inhibitor superfamily.

Plasma inter-alpha-trypsin inhibitor (I alpha TI) is a major representative of the superfamily of Kunitz-type protease inhibitors in mammals. Formerly, I alpha TI was considered to be a single polypeptide, but recent molecular genetic studies have demonstrated an unexpected multipolypeptide chain structure. The newly discovered genes and gene products form the basis of a new family of I alpha TI-related protease inhibitors.     

Specificity of the serine protease inhibitor, phenylmethylsulfonyl fluoride.

Clarified cell-free extracts were prepared from rapidly dividing $\text{Bacillus subtilis}$ cells and from rabbit liver cells. These extracts were treated with [³H]-phenylmethylsulfonyl fluoride (PMSF) and analyzed by electrophoresis in isoelectric focusing polyacrylamide gels or detergent gels. Not less than 14 proteins in the $\text{B. subtilis}$ extracts and not less than 15 proteins in rabbit liver extracts reacted covalently with PMSF. These results suggest that PMSF is not as specific for serine proteases as sometimes supposed, and its effects in physiological experiments should be interpreted with caution.     

Identification of two structurally related proteins involved in proteolytic processing of precursors targeted to the chloroplast.

Two proteins of 145 and 143 kDa were identified in pea which co-purify with a chloroplast processing activity that cleaves the precursor for the major light-harvesting chlorophyll binding protein (preLHCP). Antiserum generated against the 145/143 kDa doublet recognizes only these two polypeptides in a chloroplast soluble extract. In immunodepletion experiments the antiserum removed the doublet, and there was a concomitant loss of cleavage of preLHCP as well as of precursors for the small subunit of Rubisco and the acyl carrier protein. The 145 and 143 kDa proteins co-eluted in parallel with the peak of processing activity during all fractionation procedures, but they were not detectable as a homo- or heterodimeric complex. The 145 and 143 kDa proteins were used separately to affinity purify immunoglobulins; each preparation recognized both polypeptides, indicating that they are antigenically related. Wheat chloroplasts contain a soluble species similar in size to the 145/143 kDa doublet.     

Antifibrinolytic role of a bee venom serine protease inhibitor that acts as a plasmin inhibitor

Bee venom is a rich source of pharmacologically active substances. In this study, we identified a bumblebee (Bombus ignitus) venom Kunitz-type serine protease inhibitor (Bi-KTI) that acts as a plasmin inhibitor. Bi-KTI showed no detectable inhibitory effect on factor Xa, thrombin, or tissue plasminogen activator. In contrast, Bi-KTI strongly inhibited plasmin, indicating that it acts as an antifibrinolytic agent; however, this inhibitory ability was two-fold weaker than that of aprotinin. The fibrin(ogen)olytic activities of B. ignitus venom serine protease (Bi-VSP) and plasmin in the presence of Bi-KTI indicate that Bi-KTI targets plasmin more specifically than Bi-VSP. These findings demonstrate a novel mechanism by which bumblebee venom affects the hemostatic system through the antifibrinolytic activity of Bi-KTI and through Bi-VSP-mediated fibrin(ogen)olytic activities, raising interest in Bi-KTI and Bi-VSP as potential clinical agents.     

X-ray scattering study of hagfish protease inhibitor, a protein structurally related to complement and alpha 2-macroglobulin.

A protease inhibitor from hagfish blood plasma, homologous to human alpha 2-macroglobulin, has been studied in solution using small-angle X-ray scattering; the radius of gyration, R, was found to be 7.0 nm, the molecular weight 340,000 +/- 20,000, and the largest distance within the molecule, Dmax, 22 nm. When the inhibitor reacts with chymotrypsin, its 1:1 chymotrypsin complex is found to be more compact than the native molecule, R = 6.1 nm. A very similar conformational change is observed after the protein is reacted with methylamine. The data are consistent with models consisting of two equal elliptic cylinders with the same size as the one used as a model for the complement proteins C3 and C4 [cf. Osterberg et al. (1989) Eur. J. Biochem. 183, 507-511]. In the model for the native protein, these cylinders are arranged in an extended form, and in the one for the methylamine derivative (or chymotrypsin complex), they are closer together so that the projection of their elliptic surfaces forms an angle of about 70 degrees. These models for the hagfish protease inhibitor were expanded to models for the twice as large human alpha 2-macroglobulin using symmetry operations, and the resulting alpha 2-macroglobulin models were found to agree with those emerged from earlier studies involving electron microscopy and X-ray scattering methods.     

Two large Arabidopsis thaliana gene families are homologous to the Brassica gene superfamily that encodes pollen coat proteins and the male component of the self-incompatibility response

The male component of the self-incompatibility response in Brassica has recently been shown to be encoded by the S locus cysteine-rich gene (SCR). SCR is related, at the sequence level, to the pollen coat protein (PCP) gene family whose members encode small, cysteine-rich proteins located in the proteo-lipidic surface layer (tryphine) of Brassica pollen grains. Here we show that the Arabidopsis genome includes two large gene families with homology to SCR and to the PCP gene family, respectively. These genes are poorly predicted by gene-identification algorithms and, with few exceptions, have been missed in previous annotations. Based on sequence comparison and an analysis of the expression patterns of several members of each family, we discuss the possible functions of these genes. In particular, we consider the possibility that SCR-related genes in Arabidopsis may encode ligands for the S gene family of receptor-like kinases in this species.     

The 70-kd mammalian heat shock proteins are structurally and functionally related to the uncoating protein that releases clathrin triskelia from coated vesicles.

It is shown that in immunological, structural and functional terms the uncoating protein, which catalyses ATP-dependent dissociation of clathrin triskelia from clathrin-coated vesicles is intimately related to two major stress proteins of mammalian cells. These proteins of hitherto unknown functions have polypeptide mol. wts. of 73 kd and 72 kd, respectively. They are normal cell constituents which are synthesized in increased abundance under adverse environmental circumstances, such as non-physiological temperatures or treatment with amino acid analogues.